Hyperion Confirms Leak Of AmigaOS 3.1 Source Code

[plonoma]

Maybe it's time for them to dissolve and do other things.
On top of alternatives becoming competitives. The code has leaked.
Its over, code leak is the kiss of death, writing is on the wall.

[SystemCrasher]

Maybe it's time for them to dissolve and do other things.
On top of alternatives becoming competitives. The code has leaked.
Its over, code leak is the kiss of death, writing is on the wall.

For greedy proprietary nuts its kiss of death. For those who can leverage it, its better than any blessing. E.g. take a look on Linux. It generates (or saves) orders of magnitude more money for everyone involved than these poor bastards can dream of. Hey, proprietary nuts, it is a good day to die. Good riddance.

[SystemCrasher]

At the time, Amiga outperformed and was more stable than Windows. Of course, that was a long time ago. Personally, I think that Hyperion has a great future with porting jobs.

At the end of day, either you have MMU in HW and do MMU magic properly, or not. This is almost a binary value. Though Win9x managed to do it half-way, and so in this sense win9x and amiga OS are more or less on par: runaway or malicious program can change OS parts and ruin everything. At this point whole definition of stability is void. It is not "if" but "when". All serious OSes these days are much better at fending off wrong memory accesses and ensuring system only changed in special, tightly controlled cases (e.g. by loading kernel drivers/modules by privileged user). Linux, NT, modern MacOS, *BSDs, ... are like this. At this point we can have at least some notion of "stability".

[Chewi]

Amiga Inc really does a good job tarnishing the name of one of the best homecomputers. For those who don`t know, Amiga died in 1993 with Commodore, its been a row of license transfers since them - for now 3 decade old tech (which was once insanely great, but 3 decades is an enternity in tech).

But well, the company is dead, and the breakthroughs are forgotten or advertised from other companies/projects nowadays

I think you're a little behind on the current situation. Amiga Inc were indeed bastards who did nothing for throw the name at completely unrelated projects but this has nothing to do with them. They entered a long legal battle with Hyperion over rights to the OS. I thought they were going to win but they eventually ran out of money so Hyperion won by default, much to the relief of everybody. They've been doing a good job in what must be a tricky niche market so I feel bad that this has happened to them.

[eltomito]

Hi! I used to program for the Amiga and I must disagree with some of the things you say:

it supported long filenames, but the actual fs was so horrible that it made FAT look like a clean and efficient design;

What exactly was wrong with the filesystem? I actually liked the fact that each block on the disk had a link to the next one in the same file, so even if the disk was superficially erased (AKA the file table was deleted), you could still recover a file you deleted by accident if you managed to find at least one block that belonged to it.

it supported shared libraries but lacked a dynamic linker and so on and so on...

Well, maybe that was because most people programmed in assembly which didn't need a dynamic linker. I never even tried to find out if a C compiler existed for the Amiga. I assume if there was one, it would probably have its own dynamic linker, right?

On a funny note, I remember the concept of dynamic libraries was so alien to PC programmers that they used fixed ram addresses to call shared library functions in their programs believing that all standard libraries would always get loaded in the same place in RAM.

[gururise]

The Amiga computer was way ahead of its time. When it was released in 1985 it was the ONLY home computer to have a unix like OS that fully supported pre-emtive multitasking (something we take for granted today) a GUI and command-shell interface, 4096 color palette, 8-bit 4-channel stereo audio and high-res animation/graphics. While none of that would make a person take a 2nd look today, you have to remember what the original Amiga was competing against. The number one computer of the time period was the Commodore 64 (8-bit CPU, 16 color gfx). The Mac was just released but did not support pre-emtive multi-tasking and was BLACK & WHITE, and the IBM PC was DOS only with CGA (4-color graphics). Against the computers of the time, it was miles ahead of anything on the market.

[jacob]

What exactly was wrong with the filesystem? I actually liked the fact that each block on the disk had a link to the next one in the same file, so even if the disk was superficially erased (AKA the file table was deleted), you could still recover a file you deleted by accident if you managed to find at least one block that belonged to it.

The problem was that there was no file table or any real index of any sort, so finding empty space, resolving a file name etc. effectively meant scanning the disk until it found what it was looking for (kind of like mounting a tar archive via a loobpack device on Linux). Not only this was incredibly inefficient, it was made worse by the fact that one weakness of the original Amigas (e.i. pre-A3000) was their poor I/O performance.

But that's not even all. Due to this linked list design, the OS had to store some metadata into each allocated disk block, and it did it by putting them at the beginning of each block. This meant that when a block was read from the disk into the destination buffer, it had to be shifted by a certain number of bytes to skip the metadata. This may sound trivial today but on a 7 Mhz CPU with a 16 bit bus, it slowed things down quite considerably.

Putting it all together, you had a disk format that required enormous amounts of I/O on a machine with limited I/O bandwidth and on top of that, each disk block had to be processed separately. It kinda worked (slowly) on floppies but once hard drives began to appear, the whole thing became unusable, which is why Commodore had to introduce the marginally less awful AFFS.

Well, maybe that was because most people programmed in assembly which didn't need a dynamic linker. I never even tried to find out if a C compiler existed for the Amiga. I assume if there was one, it would probably have its own dynamic linker, right?

Of course there were C compilers. Virtually all the productivity apps for the Amiga were written in C. Even Workbench and Intuition were written in C (as opposed to the kernel, which was BCPL). Games and demos developers used almost always assembly, but then they usually did not rely on the OS at all.

And no, the C compilers didn't have their own dynamic linkers. Using a dynamic library (whichever the language) meant you had to load it and resolve your symbol names manually. In Linux terms it was exactly as if all libs, including glibc, had to be accessed through dlopen().

On a funny note, I remember the concept of dynamic libraries was so alien to PC programmers that they used fixed ram addresses to call shared library functions in their programs believing that all standard libraries would always get loaded in the same place in RAM.

It's still largely the case on Windows. Even the so-called "a.out" binary format used by early versions of Linux was like that. There is a reason for this, though: unlike the M68000, x86 never bothered to support PC-relative addressing - or, to be more precise, you had to emulate it by copying PC to another register and then calculating the destination address manually. On an architecture with only 8 GPRs (one of which being reserved as stack pointer) and a not very orthogonal instruction set, the implied overhead could be high. Try compiling some code for 32-bit x86 with the -fPIC option (Position Independent Code), disassemble the result and stare at the horror

The ELF format decided to enforce position independent code for libraries in all cases and to accept the price for it. Windows, which uses COFF, kept requiring fixed addresses for DLLs (altough it now supports PIC DLLs as well). Mercifully, amd64 introduced a proper PC-relative addressing mode like the 68000, which finally solved this problem once and for all.

[eidolon]

Amigans are a niche enthusiast community (ex. http://arstechnica.com/gaming/2015/1...ld-amiga-game/), like that of Dreamcast owners. It is also a somewhat fractured lot, as xfce_fanboy pointed out. There are likely a handful of Amiga systems still being used outside of that enthusiast community (e.g. http://woodtv.com/2015/06/11/1980s-c...s-heat-and-ac/). The Amiga brand has been passed around to the extent that it is rather worthless. And every few years a special breed of snake oil salesperson appears promising a modest resurgence in Amiga development activity, which is still more than they can possibly deliver, only to end up pushing an x86-64 Linux Mint remix (Commodore OS Vision) or Xubuntu remix (Amiga OpenLinux | Amiga Enterprise Linux | Ami/LX) instead, which was likely their intent all along. Eventually they realize there is no real money to be made, and they and their promises of code contributions fade into the background.

I have no greater expectations of AmigaOS than I do of eComStation, and I have no greater expectations of AROS than I do of Haiku. They have their place. There isn't an Amiga-like OS in this day and age that is really intended for non-enthusiast or non-legacy usage.

I don't think Hyperion Entertainment, whose 2015 was somewhat tumultuous, is too terribly concerned about the source code leak from a competitive perspective. I think this is mostly for show, a shaking of the branches to make AmigaOS (more specifically the valuation and mindshare thereof) look more imposing than it is.

Fly, AmigaOS 3.1 source code from 1994, be free.

(Reference to extinct flightless bird intentional)

[kebabbert]

It was never really great. Compared to the OSes used on micros back then, namely MS-DOS, Atari's TOS and the first versions of MacOS, it had very modern features, but their implementation often looked like a long litany of WTFs. It had multitasking, but there has never been a "proper" way to kill a process; it supported long filenames, but the actual fs was so horrible that it made FAT look like a clean and efficient design; it supported shared libraries but lacked a dynamic linker and so on and so on...

Amiga was competing against e.g. MS-DOS, a monochrome monitor with CGA graphics, that is 4 bit colors with a command line. Pit that against a multi tasking mouse driven GUI with 4096 colors.

Sure the Amiga had different non optimal implementations - but it worked. It had a implementation, whereas the competition had not. For instance, the Amiga had less than optimal multi tasking because it had no memory protection - but no commodity computer at that time even had multi tasking! Frankly, I think this is stupid complaining about how inferior Amiga is to todays Linux - of course Linux has better multi tasking today 20 years later. If you dont understand how much ahead Amiga was at that time, competing against much much inferior computers back then, you will never understand how much I am try to tell you.

I think it is really really dumb to say that MS-DOS was a better OS because AmigaOS had no memory protecting multi tasking. Well, MS-DOS had no multi tasking at all. Neither had Mac OS, Atari ST, etc etc etc. Sure, the Amiga filesystem was not as good as today, but the competition had much worse filesystems. One thing I liked with the Amiga filesystem, was it's ability to cram in 880 KB data on a 3.5" floppy, whereas x86 only fit in 720KB on the same 3.5" floppy.

You can not judge Amiga by today's standards, but you need to compare to the computers from the same time period. And if you have the IQ to do that, Amiga were totally superior in every way back then.

[jacob]

Amiga was competing against e.g. MS-DOS, a monochrome monitor with CGA graphics, that is 4 bit colors with a command line. Pit that against a multi tasking mouse driven GUI with 4096 colors.

Sure the Amiga had different non optimal implementations - but it worked. It had a implementation, whereas the competition had not. For instance, the Amiga had less than optimal multi tasking because it had no memory protection - but no commodity computer at that time even had multi tasking! Frankly, I think this is stupid complaining about how inferior Amiga is to todays Linux - of course Linux has better multi tasking today 20 years later. If you dont understand how much ahead Amiga was at that time, competing against much much inferior computers back then, you will never understand how much I am try to tell you.

I think it is really really dumb to say that MS-DOS was a better OS because AmigaOS had no memory protecting multi tasking. Well, MS-DOS had no multi tasking at all. Neither had Mac OS, Atari ST, etc etc etc. Sure, the Amiga filesystem was not as good as today, but the competition had much worse filesystems. One thing I liked with the Amiga filesystem, was it's ability to cram in 880 KB data on a 3.5" floppy, whereas x86 only fit in 720KB on the same 3.5" floppy.

You can not judge Amiga by today's standards, but you need to compare to the computers from the same time period. And if you have the IQ to do that, Amiga were totally superior in every way back then.

Re-read my post. I never said that MS-DOS was superior, of course it wasn't. The AmigaOS was way ahead of the others, absolutely, but it is also a fact that in some ways, it was misdesigned (meaning that it could, and should, have been designed differently given the hardware and technology of 1986) and in some ways it was half-baked or unfinished. It didn't have memory protection because that wasn't possible on a M68000 but the fact that there was no documented way to kill a process had nothing to do with memory protection, it was just a serious shortcoming. Sure, MS-DOS didn't have multitasking at all, so how does that make it less of a shortcoming ?

The problem with the filesystem was not that it wasn't as good as today, but that it was bad *at the time*. It's sad to say it, but yes, MS-DOS's FAT was a better filesystem. In this one specific case, Microsoft's design was indeed superior to AmigaOS's. It's interesting that you should cite the 880Kb size, because that's exactly what the origin of the issue was. Commodore believed that claiming a larger floppy capacity than competing machines would be a good selling point and, indeed, it was. Funnily enough, initially it was sometimes perceived as a greater advantage than multitasking itself. Many reviewers and magazines of the era regarded multitasking as some sort of novelty of little practical value, comments such as "yeah but at the end of the day, you really only work with one application at a time etc..." were not uncommon. On the other hand, 880Kb floppies appealed to everyone.

Unfortunately, Commodore thus set out to design a fs that would maximise floppy capacity and one of the solutions to do that was to get rid of "unnecessary" tables. The result was probably the slowest and most unreliable filesystem I had the misfortune to use.

The Amiga is a legend but like any technology way ahead of its time, it used some brilliant ideas and some that turned out to be dead ends. Acknowledging the latter is not bashing, it is how we learn to make better things. Ultimately, I would argue that it is the difference between true love and blind fanboyism.